datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
shogi_hao_depth9
Summary
Training Data for Shogi AI Development
Contents
kifu.tag=train.depth=9.num_positions=1000000000.start_time=??????????.thread_index=???.bin ... Training Data
The training data are provided in the YaneuraOu PackedSfenValue format.
This dataset was generated using Hao with a search depth of 9.
The training data have not been shuffled. It is recommended to shuffle the training data before use. Additionally, positions within this dataset have not been replaced with… See the full description on the dataset page: https://huggingface.co/datasets/nodchip/shogi_hao_depth9.shogi_suisho5_depth9
Summary
Training and Validation Data for Shogi AI Development
Contents
shuffled.7z.00? ... Training Data
shuffled.bin ... Validation Data
The training and validation data are in the YaneuraOu PackedSfenValue format.
Both datasets were generated using Suisho5 with a search depth of 9.
The training and validation data have already been shuffled.
Positions within these datasets have been replaced with the PV (Principal Variation) leaf node from the quiescence search of the… See the full description on the dataset page: https://huggingface.co/datasets/nodchip/shogi_suisho5_depth9.shogi-balanced-positions-2025-extended
Shogi Balanced Positions 2025, extended to move 37
Balanced shogi (Japanese chess) opening positions, from move 32 to move 37.
The starting point is yaneurao's Balanced Position Collection 2025, which
contains 26,273 positions at move 32, each verified balanced by a 200M-node Suisho10
search (|eval| ≤ 50 centipawns). That collection stops there. This dataset walks it
five plies further, keeping every position balanced, and ends with 6,045,847
positions at move 37.
It comes in… See the full description on the dataset page: https://huggingface.co/datasets/qleap/shogi-balanced-positions-2025-extended.shogi_suisho5_depth9_entering_king
Summary
Training Data for Shogi AI Development
Contents
kifu.tag=suisho5.entering_king.depth=9.num_positions=500000000.start_time=??????????.thread_index=???.bin ... Training Data
The training data are provided in the YaneuraOu PackedSfenValue format.
This dataset was generated using Suisho5 with a search depth of 9. The starting positions were selectively chosen from Floodgate games spanning the years 2015 to 2024, specifically focusing on scenarios involving entering… See the full description on the dataset page: https://huggingface.co/datasets/nodchip/shogi_suisho5_depth9_entering_king.eval-shogi-testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 6,
"total_frames": 4008,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:6"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/aoiro23467/eval-shogi-test.shogi-testThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 28,
"total_frames": 26510,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:28"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/aoiro23467/shogi-test.shogi_hao_depth9_Shuffled_qsearch